[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

US8064991B2 - Method of fetal and maternal ECG identification across multiple EPOCHS - Google Patents

Method of fetal and maternal ECG identification across multiple EPOCHS Download PDF

Info

Publication number
US8064991B2
US8064991B2 US11/970,553 US97055308A US8064991B2 US 8064991 B2 US8064991 B2 US 8064991B2 US 97055308 A US97055308 A US 97055308A US 8064991 B2 US8064991 B2 US 8064991B2
Authority
US
United States
Prior art keywords
maternal
ica
fetal
signal
ica output
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related, expires
Application number
US11/970,553
Other languages
English (en)
Other versions
US20090177101A1 (en
Inventor
Lawrence T. Hersh
Sai Kolluri
Bruce A. Friedman
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
General Electric Co
Original Assignee
General Electric Co
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by General Electric Co filed Critical General Electric Co
Priority to US11/970,553 priority Critical patent/US8064991B2/en
Assigned to THE GENERAL ELECTRIC COMPANY reassignment THE GENERAL ELECTRIC COMPANY ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: FRIEDMAN, BRUCE A., HERSH, LAWRENCE T., KOLLURI, SAI
Priority to GB0823016.1A priority patent/GB2456375B/en
Priority to JP2009000847A priority patent/JP5271718B2/ja
Priority to DE102009003317.3A priority patent/DE102009003317B4/de
Publication of US20090177101A1 publication Critical patent/US20090177101A1/en
Application granted granted Critical
Publication of US8064991B2 publication Critical patent/US8064991B2/en
Expired - Fee Related legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A61B5/02411—Detecting, measuring or recording pulse rate or heart rate of foetuses
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/024—Detecting, measuring or recording pulse rate or heart rate
    • A61B5/0245—Detecting, measuring or recording pulse rate or heart rate by using sensing means generating electric signals, i.e. ECG signals
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/344—Foetal cardiography
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A61B5/349—Detecting specific parameters of the electrocardiograph cycle
    • A61B5/366—Detecting abnormal QRS complex, e.g. widening
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/389—Electromyography [EMG]
    • A61B5/391—Electromyography [EMG] of genito-urinary organs
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/43—Detecting, measuring or recording for evaluating the reproductive systems
    • A61B5/4306—Detecting, measuring or recording for evaluating the reproductive systems for evaluating the female reproductive systems, e.g. gynaecological evaluations
    • A61B5/4343—Pregnancy and labour monitoring, e.g. for labour onset detection
    • A61B5/4362—Assessing foetal parameters
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/321—Accessories or supplementary instruments therefor, e.g. cord hangers
    • A61B5/322—Physical templates or devices for measuring ECG waveforms, e.g. electrocardiograph rulers or calipers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Definitions

  • the present disclosure generally relates to a method of non-invasively monitoring the heartbeat and ECG of an unborn fetus. More specifically, the present disclosure relates to a method of determining maternal and fetal heart rates and selecting maternal and fetal ECG signals from an electrocardiogram (ECG) obtained from a maternal patient during continuous monitoring of the maternal patient.
  • ECG electrocardiogram
  • ECG electrocardiogram
  • obtaining an accurate fetal ECG is difficult due to the weaker fetal information obtained from the abdomen of the mother.
  • the mother's ECG is present and is usually significantly larger than the ECG of the fetus.
  • uterine contractions may be present, which result in large electrical artifacts that obliterate or mask the fetal signal.
  • the mother is experiencing discomfort and is unable to lie still, which creates large electrical muscle artifacts.
  • ICA independent component analysis
  • the output from the ICA algorithm can be used to provide a signal for finding the fetal heart rate.
  • one channel of the multi-channel abdominal maternal ECG leads is used for determining the fetal heart rate.
  • the ICA algorithm is performed on a set of waveforms over an epoch having a determined length, such as 4 to 5 seconds.
  • the channel jumping of the ICA output is a problem since a separated fECG signal may not be in the same waveform position from epoch to epoch.
  • automated techniques for identifying and monitoring the fetal ECG are further complicated by the changing channels of both the fetal and maternal signals over multiple epochs. Therefore, it becomes necessary to have a method and means to recognize which among the scrambled plurality of ICA output waveforms from any epoch is a fetal ECG, a maternal ECG, a uterine contraction or just noise. A need exists to make this determination for each epoch of a series of epochs such that the fetal ECG and maternal ECG can be monitored over an extended period of time.
  • the present disclosure generally relates to a method of monitoring maternal and fetal vital signs, including ECG and heart rate information, obtained from a maternal patient. More specifically, the present disclosure relates to a method of identifying which ICA channels derived from ECG signals from the maternal patient are produced by fetal and material ECG sources as the algorithms moves through a series of sequential epochs.
  • a plurality of ECG electrodes is placed on the abdomen of the maternal patient to obtain ECG signals from the maternal patient.
  • the electrodes are connected to an acquisition system such that input ECG waveforms are detected and received for monitoring heart rate or other ECG properties.
  • an independent component analysis (ICA) algorithm is applied to each of the waveforms separately over a defined epoch.
  • the defined epoch is between 4 and 5 seconds in length such that the ICA algorithm is applied to the waveform over the entire duration of the epoch.
  • the ICA algorithms are a group of well-known and widely available processing algorithms.
  • an ICA output waveform is generated and associated with each of the plurality of channels.
  • the ICA algorithm acts as a filtering and cleaning algorithm that enhances and identifies hidden independent sources from the original input waveforms from the ECG electrodes and generates more useful output waveforms.
  • the ICA algorithm generates the same number of separate, individual ICA output waveforms.
  • the present disclosure provides improved processing techniques for identifying which channels of the plurality of ICA output channels are fetal or maternal signal sources for the current epoch. Since the maternal and fetal ICA output signals can change channels from one epoch to the next, the system and method of the present disclosure operates to identify the channels that are either the maternal signal or the fetal signal for each individual epoch.
  • a discrete Fourier transform is computed using a fast Fourier transformer (FFT) for each ICA output waveform.
  • FFT fast Fourier transformer
  • the FFT algorithm is well known and will be used whenever a DFT is required for the methods being disclosed in the following.
  • the FFT for each of the ICA output waveforms is classified and the significant frequency peaks and the location of such peaks are determined for the ICA output waveform for each of the plurality of channels.
  • the system compares the peaks of the FFT for each of the ICA output waveforms to a known, typical maternal signal determined from a previous epoch. If the frequency peaks match the maternal signal from the previous epoch, the ICA output waveform is classified as being a maternal signal and is stored for further processing.
  • the system determines whether the FFT for the ICA output waveform matches a known fetal signal from a previous epoch. If the signal matches the known fetal signal from a previous epoch, the ICA output waveform is classified as being a fetal signal and is stored in memory for further processing.
  • the system can compare the FFT for the ICA output waveform to a known uterine signal. If the FFT for the ICA output waveform matches the known uterine signal, the waveform is stored as a waveform representing uterine activity.
  • the ICA output waveform does not match the known maternal signal, the known fetal signal or the known uterine signal, the ICA output waveform is classified as noise.
  • the waveforms from these channels are fed to separate fetal and maternal ECG processors for analysis in a known manner. This method is carried out during each epoch such that the system and method identifies which ICA output waveforms are fetal or maternal for each epoch.
  • the system and method of the present disclosure utilizes a correlation function to determine which ICA output channels are the fetal or maternal source signals.
  • the system first obtains a fetal QRS template from some previous time, such as a previous epoch. Since a given fetal QRS template will most likely correlate with the QRS structure for a fetal source signal from an ICA output waveform of the present epoch, the system calculates a correlation of the known fetal QRS template through time across the waveforms of the current epoch.
  • the correlation of the channel including the fetal properties will result in high correlation as the template aligns with the QRS of the ICA output waveform, which implies that there will be high correlation peaks spaced in a regular pattern if the particular ICA waveform is a fetal source signal. Based upon the ICA output signal that generates the best correlation, this ICA output waveform is identified as including the fetal signal. If the remaining ICA output waveform channels are not a fetal source signal, these channels will have a lower, more varied correlation signal and will not be classified as a fetal source signal.
  • the system and method can utilize a maternal QRS template that is also determined from some previous time period, such as a previous epoch.
  • a maternal QRS template that is also determined from some previous time period, such as a previous epoch.
  • the system calculates a correlation for each of the ICA output waveforms and, based upon the correlation, determines which of the ICA output waveforms is a maternal source signal.
  • the waveforms from these channels are directed to a fetal ECG processor and a maternal ECG processor for further processing and display of the ECG signal for both the patient and the fetus.
  • the two methods discussed namely the FFT and the correlation methods, each have their own advantages and disadvantages.
  • the FFT technique requires no template, is capable of immediately identifying the fetal and maternal heart rates by noting the position of fundamental and harmonic peaks in the FFTs of the appropriate source signal waveforms, and allows easy frequency domain filtering to aid the system in heart rate calculations.
  • the correlation technique aids in heart rate calculations by producing clear correlation peaks with the needed period, but requires a template that may not always be easily available.
  • FIG. 1 is an illustration of a set of ECG electrodes positioned on an abdomen of a maternal patient in accordance with the present disclosure
  • FIG. 2 is a data flow diagram of the method of the disclosure that utilizes an ICA algorithm applied to fetal ECG extraction;
  • FIG. 3 is an example of ICA output waveforms from electrodes applied to a maternal abdomen
  • FIG. 4 is an example of the frequency content of a fetal waveform after an FFT
  • FIG. 6 is a sample fetal QRS template
  • FIG. 7 is an example correlation output using the fetal QRS template on a fetal ICA output waveform
  • FIG. 8 is an example correlation output utilizing the fetal QRS template on a maternal waveform
  • FIG. 9 is an example correlation output utilizing a fetal QRS template on a noise separated waveform
  • FIG. 10 is a flowchart illustrating the steps to determine whether a waveform includes the fetal signal or the maternal signal using frequency content.
  • FIG. 11 is a flowchart for deciding whether a waveform includes the fetal signal utilizing a correlation analysis technique.
  • the abdominal ECG electrodes 18 should not be placed too close together and should involve a wide coverage of the abdomen.
  • the regularly spaced sixteen electrodes are spaced over the entire skin surface of the patient 10 .
  • Each of the ECG electrodes 18 detects electrical signals present on the skin of the patient and returns the sensed electrical signals to the ECG monitor 12 over the series of separate patient leads 16 .
  • the ECG monitor 12 receives a set of multiple input waveforms from the ECG electrodes 18 .
  • the ECG monitor 12 includes operating programs and software, to be described in much greater detail below, that operate to separate ECG and physiological information relating to the mother from ECG information and physiological parameters resulting from the fetus.
  • the physiological information that may be derived for both the mother and the fetus include at least maternal and fetal heart rate, and may include maternal respiratory rate, other maternal and fetal ECG characteristics and the electrohistogram (EHG).
  • the initial step in the method of the present disclosure is to prepare the abdomen of a pregnant patient for the application of the array of electrodes, as illustrated in step 24 .
  • This preparation typically includes removing clothing covering the abdomen 20 and cleaning the abdomen to prepare the skin for good surface contact with the individual ECG electrodes 18 , as shown in FIG. 1 .
  • the next step in the method is to place the sixteen ECG electrodes, the one reference ECG electrode and the one ground ECG electrode on the patient, as set forth in step 26 .
  • each of the electrodes adheres to the patient's abdomen and creates the required low impedance skin to electrode contact to sense electrical signals present on the surface of the patient due to electrical activity within both the patient and the fetus.
  • the ECG monitor 12 includes front end electronics, differential amplifiers, isolation devices and common mode rejection components that receive the individual input waveforms from the electrodes and provide for initial processing of the sixteen separate input waveforms received on the sixteen separate channels at the ECG monitor 12 .
  • the ECG monitor includes computing and storage means for receiving the individual input waveforms and outputting monitoring data for use by a physician.
  • Such digital instrumentation can process the input waveforms by applying algorithms and filtering operations.
  • the computing means can include storage components for recording the input waveforms, as will be described in detail below.
  • the ECG monitor 12 provides initial processing of the input waveforms received across the sixteen separate channels and records the input waveform in memory contained within the ECG monitor. Since the waveform for each channel is received continuously at the ECG monitor, the ECG monitor 12 stores the input waveforms continuously in a memory device contained within the ECG monitor.
  • the input waveforms obtained from the sixteen channels connected to the ECG monitor 12 include a large amount of noise and unwanted signal information.
  • the system and method shown in FIG. 2 utilizes blind source separation (BSS) using an independent component analysis (ICA) algorithm applied to the sixteen separate channels from the individual electrodes applied to the patient's abdomen.
  • BSS blind source separation
  • ICA independent component analysis
  • the ECG monitor stores the continuous input waveforms from each of the sixteen channels over a defined period of time.
  • the stored input waveforms for each of the sixteen channels connected to the ECG electrodes are stored and segmented into individual epochs.
  • an epoch has a fixed duration such that signal processing techniques can be carried out on each of the individual waveforms over the epoch.
  • an epoch typically has a duration of approximately 4-5 seconds, although other durations are contemplated. It is preferred that each epoch have a duration that is greater than the heartbeat period for both the patient 10 and the fetus.
  • An epoch duration of 5 seconds is shown in the embodiment of the present disclosure, as illustrated by the epoch duration 32 shown in FIG. 3 .
  • the ICA algorithm is applied to the set of the input waveforms at step 30 to generate clean and separated signals that represent the electrical activity of the mother, the fetus and any other independent sources that may be present during the acquisition of the sixteen ECG channels 18 .
  • the BSS/ICA data processing techniques of step 30 are well known to those of skill in the art and are readily available from numerous outlets or can be developed and optimized for an intended purpose.
  • the ICA algorithm of step 30 is preferably implemented in real time on a computing means contained within the ECG monitor.
  • the ICA algorithm filters the input waveforms obtained directly from the ECG electrodes such that the filtered waveforms can be easily further processed by other components.
  • ICA algorithms are well known and have been used for quite some time.
  • the ICA algorithm could be any known algorithm such as FASTICA or CUBICA, although other types of ICA algorithms are contemplated as being within the scope of the present disclosure.
  • the ICA algorithm 30 Since each of the input waveforms is received on a separate channel from a separate ECG electrode, the ICA algorithm 30 generates sixteen separate waveforms that are each related to separate, independent sources, as indicated in step 34 . Since the ICA algorithm 30 filters and removes much of the noise from the input waveforms, the sixteen ICA output waveforms, generated in step 34 , can be utilized to determine on which channel the fetal and maternal heart rate and ECG signals are present. Once the channels have been identified, further processing can be conducted on the maternal and fetal signals.
  • the ICA output waveforms 36 that are generated from the ICA algorithm shown in step 30 of FIG. 2 .
  • the ICA output waveforms 36 are present along each of the sixteen channels 38 , where each channel 38 represents one of the plurality of waveforms obtained from the electrodes placed on the patient's abdomen.
  • many of the sixteen channels 38 shown in FIG. 3 include ICA output waveforms that appear to present only noise or signals having very little usefulness in monitoring the fetal and maternal ECGs and heart rates.
  • the ICA output waveform on channel 8 includes a series of signature QRS events 40 that represent the heartbeat and ECG signal from the fetus.
  • the fetal signal 42 for the epoch 32 is said to reside on channel 8 .
  • channel 15 includes a series of QRS events 42 that represent the heart rate and ECG signal received from the maternal patient.
  • channel 15 from the epoch 32 includes the maternal signal 46 .
  • the material signal 46 also includes T-waves following the application of the ICA algorithm.
  • the ICA output waveforms 36 can be utilized to determine significant physiological properties for both the patient and the fetus. Specifically, the ICA output waveform 36 present on channel 15 can be utilized to determine physiological properties from the maternal patient, while the ICA output waveform present on channel 8 can be utilized to determine physiological properties for the fetus.
  • This use of the ICA algorithm to generate the ICA output waveforms shown in FIG. 3 is generally well known and defined in the state of the art.
  • the fetal signal 42 and the maternal signal 46 can, and typically do, change channels from one epoch to another, such that although the maternal and fetal signals may be determined for a first epoch, the maternal and fetal signals for the next epoch could be located on different channels.
  • the channel including both the fetal signal 42 and the maternal signal 46 must be known, which presents a significant challenge addressed by the present disclosure.
  • the next step in the method of the present disclosure is to utilize some type of waveform sorting algorithm 48 to identify which channels include the maternal source signal and the fetal source signal.
  • the waveform sorting algorithm 48 determines which channels include the maternal and fetal signals
  • the information from these channels are fed to either a maternal ECG processor 50 or a fetal ECG processor 52 for further processing.
  • the maternal ECG processor 50 can determine physiological properties for the maternal patient based upon analysis of the ICA output waveform that includes the maternal signal, while the fetal ECG processor 52 can perform processing techniques to generate physiological properties for the fetus.
  • the system can also include a uterine activity processor 54 that can be used to monitor the uterine activity of the patient, such as contraction strength, contraction intervals and other relevant information from the maternal patient during the contractions preceding the delivery of the fetus.
  • a uterine activity processor 54 can be used to monitor the uterine activity of the patient, such as contraction strength, contraction intervals and other relevant information from the maternal patient during the contractions preceding the delivery of the fetus.
  • one significant problem that arises by utilizing only the ICA algorithm 30 to determine which channel includes a fetal signal and a maternal signal is that, although the location of these signals can be determined for a selected epoch, the location of these signals may change from one epoch to the next.
  • at least two alternative ways of deciding which signals are on which channels have been developed in accordance with the present disclosure.
  • the first method of deciding which channels include the maternal signal and the fetal signal is generally shown and described in the flowchart of FIG. 10 .
  • the system begins the process of identifying the waveforms in step 51 .
  • the method first calculates a fast Fourier transform (FFT) for each of the ICA output waveforms, as shown in step 53 .
  • FFT fast Fourier transform
  • the method calculates each of the ICA waveforms sequentially from channel 1 to channel 16 , as will be described.
  • FIGS. 4 and 5 illustrate the FFT of a fetal waveform 56 and an FFT for a maternal waveform 58 .
  • FIGS. 4 and 5 illustrate the FFT for the fetal waveform and the maternal waveform, it should be understood that the FFT for ICA output waveforms that include neither the fetal waveform nor the maternal waveform present a different energy characteristic than those shown in FIGS. 4 and 5 .
  • the first peak 60 for the fetal waveform represents the fetal heart rate for the fetus.
  • the peak 60 occurs at approximately 2.5 Hz, which is representative of the heart rate of the fetus.
  • the maternal waveform 58 shown in FIG. 5 also includes a peak 62 that represents the maternal heart rate.
  • the peak 62 occurs at approximately 1.5 Hz, which is representative of the fundamental heart rate of the maternal patient.
  • the system determines whether all of the ICA output waveforms have been sorted in step 64 . If not all of the ICA output waveforms have been sorted, the system proceeds to step 66 and attempts to classify the FFT based upon the ICA output waveform from the current channel being analyzed. In step 66 , the system compares the energy peaks identified in step 55 to peaks that are present in a typical maternal signal calculated from a previous epoch or previously stored in a memory location. If the system determines in step 68 that the frequency peaks of the FFT waveform being analyzed correspond to a maternal signal, the system places the waveform in a maternal memory storage location for further processing, as indicated in step 70 .
  • a significant number of energy peaks occur within the frequency range 72 which, in the embodiment shown, occurs approximately between 10-20 Hz. Since the waveform 58 includes a significant amount of energy and peaks within the frequency range 72 , the waveform 58 shown in FIG. 5 is classified by the present method as a maternal waveform. In contrast to the waveform shown in FIG. 5 , the waveform 56 shown in FIG. 4 does not include many peaks within the frequency range of 10-20 Hz. Thus, the waveform 56 shown in FIG. 4 is clearly not a maternal waveform. Note that it is expected that a fetal ECG signal will contain faster components, i.e. higher energy at higher harmonic frequencies, than the maternal ECG signal.
  • step 68 when the system determines in step 68 that the frequency peaks are not maternal, the system then compares the frequency peaks to a typical fetal signal determined from a previous epoch, in step 74 . Based upon the comparison in step 74 , the system determines in step 76 whether the frequency peaks match the typical fetal signal. If the frequency peaks match the typical fetal signal, the system places the waveform in a memory storage for further processing, as shown in step 78 .
  • the waveform 56 shown in FIG. 4 has a significant amount of energy and peaks within the frequency range 80 , which is typically representative of a fetal waveform.
  • the frequency range 80 is approximately 20-50 Hz.
  • the waveform 58 shown in FIG. 5 includes a very small amount of energy in the frequency range between 20-50 Hz, which indicates that the waveform 58 is not a fetal waveform but, as previously described, the waveform 58 corresponds to a maternal waveform.
  • step 76 If the system determines in step 76 that the frequency peaks are not fetal, the system then compares the peaks to typical uterine signals determined from a previous epoch, as indicated in step 82 .
  • the typical uterine signals from the previous epoch are due to contractions of the patient's abdominal muscles, such as during a contraction. If the frequency peaks are classified as uterine in nature in step 84 , the system places the waveform in memory for further processing in step 86 . However, if the system determines that the frequency peaks are not maternal, fetal or uterine in nature, the system classifies the waveform as noise, as indicated in step 88 .
  • step 54 the system returns to step 54 to identify the significant frequency peaks for the next waveform on the next channel.
  • step 90 determines the fetal and maternal heart rate information based upon the knowledge of which channel is a fetal signal and which channel is a maternal signal, as determined in the previous analysis.
  • the system determines which channels of the plurality of channels are fetal and which are maternal
  • the fetal and maternal signals are sent to their respective fetal ECG processor 52 and maternal ECG processor 50 , as shown in FIG. 2 .
  • the method illustrated in FIG. 10 allows the system to determine which channels are fetal or maternal in characteristics so that additional processing can be performed in a manner well known to those of ordinary skill in the art.
  • the system determines which channels are fetal and maternal signals based upon an analysis and comparison of an FFT for each ICA output waveform. For example, the peaks in the FFT that were known from the previous epoch for a maternal signal and fetal signal would be expected to be present for a newly obtained epoch. Additionally, an FFT peak is expected in the neighborhood of a known heart rate for either a maternal or fetal signal, which can help distinguish the proper classification of signals.
  • the system saves the waveform for the maternal and fetal signals in memory such that the saved signals may be utilized when analyzing the ICA output waveforms for the next epoch. In this manner, the system and method of the present disclosure utilizes information from the most recent epoch to classify signals for the current epoch.
  • the system can also utilize a correlation method to classify which channels of the ICA output waveforms are the fetal source signals or the maternal source signals.
  • Algorithms for calculating correlations are well known by those skilled in the art and are essentially a measure of the agreement of the waveform pattern between two input signals.
  • the correlation algorithm requires a known template as one of the inputs. Using a time interval around some benchmark event such as a QRS event or a uterine contraction event can specify the template for use in the correlation calculation, but other means can be used to find an appropriate template.
  • the system starts the identification process in step 92 .
  • the system obtains a fetal QRS template from a previous time period or a previous epoch, as illustrated in step 93 .
  • the fetal QRS template 94 defines a QRS episode that includes a Q peak 96 and an R peak 98 over a time duration 100 .
  • the QRS template 94 is determined during a previous epoch based upon the identification of which ICA output waveform includes the fetal signal.
  • the system calculates a correlation for each of the ICA output waveforms based upon the QRS template 94 , as illustrated in step 102 .
  • the correlation of the ICA output waveform on each channel of the plurality of channels with the fetal QRS template 94 from a previous epoch will result in the waveform not having a high correlation if the waveform does not include the fetal signal, while the QRS template, when run across a waveform including the fetal signal, will give high correlation peaks with the period of the fetal heart rate.
  • the waveform that generates high correlation peaks will be the waveform that is a fetal source signal, while those waveforms that result in low correlation will be waveforms that do not include fetal components.
  • the correlation signal 104 includes a number of peaks 106 that extend above a correlation baseline 108 .
  • the threshold 108 is positioned at a correlation of 0.8, which indicates a high degree of correlation between the template and the signal being analyzed.
  • the correlation signal 104 includes numerous periodic peaks 106 that extend above the threshold 108 .
  • the correlation signal 104 shown in FIG. 7 has a very high correlation, indicating that the channel being analyzed is fetal in its characteristics.
  • FIG. 8 thereshown is a correlation signal 110 for a different channel than shown in FIG. 7 .
  • the correlation signal 104 clearly indicates a fetal signal
  • the correlation signal 110 shown in FIG. 8 does not include the fetal signal.
  • the ICA output waveform on the channel shown in FIG. 8 was determined to be a maternal signal.
  • the correlation signal 110 shown in FIG. 8 does not represent the fetal signal, but instead represents the maternal signal.
  • FIG. 9 shows the correlation signal 112 for an ICA output waveform on yet another channel that represents noise. As clearly shown, none of the peaks 106 reach the threshold 108 such that the ICA output waveform on the channel being analyzed clearly does not include the fetal signal.
  • the system determines in step 114 whether all of the waveforms on all of the channels have been evaluated. If not all of the waveforms have been evaluated, the system picks a channel in step 116 and evaluates the waveform correlation. If the correlation signal for the channel has the highest appropriately spaced peaks, as determined in step 118 , the system determines that the channel includes the fetal signal and places the ICA output waveform in memory storage for further processing, as indicated in step 120 . However, if the waveform does not have the appropriately spaced peaks, the system returns to step 114 and continues to analyze each channel to determine which of the channels includes the fetal signal. The method shown in FIG. 11 is operable to determine which of the channels includes the fetal signal based upon a correlation between the ICA output waveforms and the fetal QRS template shown in FIG. 6 .
  • the correlation method was described as identifying which channel includes the fetal signal, a similar method is carried out utilizing a maternal QRS template to determine which of the channels includes the maternal signal.
  • the system generates a correlation of the ICA output waveform for each channel relative to the maternal QRS template. Based upon which channel includes the highest and appropriately spaced peaks, the system determines which channel includes the maternal signal. Similar steps can be carried out to determine which channel includes signals related to uterine activity.
  • step 122 which corresponds to step 48 in FIG. 2 , the system then transfers the ICA output waveforms from the required channels to the fetal ECG processor 52 , the maternal ECG processor 50 and the uterine activity processor 54 for additional analysis.
  • the method in FIG. 11 relies upon a fetal QRS template that was obtained from a previous epoch.
  • a maternal QRS template from a previous epoch is utilized to identify which processing channel includes the maternal signal.
  • an issue arises when the first epoch is being analyzed by the system and method of the present disclosure. On the first epoch, a decision must be made as to which waveforms are fetal, maternal, uterine, muscle artifacts or some other artifact, since the QRS template from a past epoch is not available. Once this determination is made, the system can utilize the QRS template from the past epoch to carry out the method described.
  • the initial decision as to which channel includes the various signals can be based upon other criteria, such as the rate at which the QRS's occur, the frequency content of the QRS complexes, the amplitude of the signal, and the general frequency content of the signals.
  • the fetal QRS complex is typically known to have a slightly higher frequency content than the maternal QRS and the fetal heart rates are higher than maternal heart rates.
  • various different methods to generate the first determination of which channel includes the maternal and fetal signals can be utilized while operating within the scope of the present disclosure. Once the first fetal and maternal signals have been determined, the system can utilize the methods described above to continue to monitor which channel includes the maternal signal and the fetal signal.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Cardiology (AREA)
  • Engineering & Computer Science (AREA)
  • Animal Behavior & Ethology (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Pathology (AREA)
  • Gynecology & Obstetrics (AREA)
  • Pregnancy & Childbirth (AREA)
  • Physiology (AREA)
  • Urology & Nephrology (AREA)
  • Pediatric Medicine (AREA)
  • Reproductive Health (AREA)
  • Signal Processing (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)
US11/970,553 2008-01-08 2008-01-08 Method of fetal and maternal ECG identification across multiple EPOCHS Expired - Fee Related US8064991B2 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
US11/970,553 US8064991B2 (en) 2008-01-08 2008-01-08 Method of fetal and maternal ECG identification across multiple EPOCHS
GB0823016.1A GB2456375B (en) 2008-01-08 2008-12-18 Method of fetal and maternal ECG identification over multiple epochs
JP2009000847A JP5271718B2 (ja) 2008-01-08 2009-01-06 複数の時間区分にわたり胎児及び産婦のｅｃｇを特定する方法
DE102009003317.3A DE102009003317B4 (de) 2008-01-08 2009-01-07 Verfahren zum Überwachen von mütterlichen und fötalen Vitalzeichen einer Mutter und eines Fötus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US11/970,553 US8064991B2 (en) 2008-01-08 2008-01-08 Method of fetal and maternal ECG identification across multiple EPOCHS

Publications (2)

Publication Number Publication Date
US20090177101A1 US20090177101A1 (en) 2009-07-09
US8064991B2 true US8064991B2 (en) 2011-11-22

Family

ID=40343750

Family Applications (1)

Application Number Title Priority Date Filing Date
US11/970,553 Expired - Fee Related US8064991B2 (en) 2008-01-08 2008-01-08 Method of fetal and maternal ECG identification across multiple EPOCHS

Country Status (4)

Country Link
US (1) US8064991B2 (ja)
JP (1) JP5271718B2 (ja)
DE (1) DE102009003317B4 (ja)
GB (1) GB2456375B (ja)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8892181B2 (en) 2011-10-21 2014-11-18 Mindchild Medical, Inc. Non-invasive fetal monitoring
US8897862B2 (en) 2011-01-25 2014-11-25 Public University Corporation Nara Medical University Fetal cardiac potential signal extraction program, fetal cardiac potential signal discriminating apparatus, and pregnancy monitoring system using the same
WO2015038572A1 (en) * 2013-09-10 2015-03-19 Guerrero Juan R Diagnostic digital data mining of biological waves
US20150335886A1 (en) * 2012-11-26 2015-11-26 Guangzhou Sunray Medical Apparatus Co., Ltd. Bio-feedback type device for physical pain relief at childbirth
US11224402B2 (en) 2017-09-13 2022-01-18 Koninklijke Philips N.V. Fetal heart rate detection method to discriminate from other periodic signal

Families Citing this family (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7333850B2 (en) * 2004-05-28 2008-02-19 University Of Florida Research Foundation, Inc. Maternal-fetal monitoring system
AU2005280443B2 (en) 2004-07-30 2011-02-03 Avent, Inc. Antimicrobial silver compositions
EP1814438B8 (en) * 2004-11-08 2009-04-01 Koninklijke Philips Electronics N.V. Safe identification and association of wireless sensors
JP5509437B2 (ja) * 2010-03-01 2014-06-04 国立大学法人山口大学 超音波診断装置
CN102525452B (zh) * 2011-10-19 2014-11-12 广东工业大学 一种基于斜投影的单通道胎儿心电盲分离装置及分离方法
CN103845079B (zh) * 2014-02-14 2015-12-30 广东工业大学 一种基于盲分离的多普勒胎心音瞬时心率的检测方法
CN103876731B (zh) * 2014-03-25 2016-01-13 电子科技大学 一种胎儿心电信号提取装置及方法
DE102014013710A1 (de) 2014-09-16 2016-03-17 Personal Medsystems Gmbh Verfahren zum Erzeugen eines akustischen Signals und Elektrokardiogrammgerät
WO2016091984A1 (en) * 2014-12-12 2016-06-16 Koninklijke Philips N.V. Device and method for measuring a physiological characteristic of a subject
CN104545893B (zh) * 2015-01-12 2018-01-09 南京大学 对分离的胎儿心电图中的qrs波的真伪进行辨识的方法
WO2016173639A1 (en) * 2015-04-29 2016-11-03 Brainlab Ag Detection of the heartbeat in cranial accelerometer data using independent component analysis
JP6649982B2 (ja) * 2018-05-10 2020-02-19 アトムメディカル株式会社 胎児心電信号処理方法及び胎児心電信号処理装置
JP7155939B2 (ja) 2018-11-26 2022-10-19 住友金属鉱山株式会社 内部電極用ペーストの製造方法
CN111434305A (zh) * 2019-01-15 2020-07-21 中山大学 基于卷积编解码神经网络的胎儿心电提取系统及方法
JP7418740B2 (ja) * 2019-02-21 2024-01-22 株式会社クラウドセンス 胎児心拍監視システム
JP2020138022A (ja) * 2019-02-27 2020-09-03 国立大学法人信州大学 胎児心拍数測定方法並びに胎児および母体の同時監視方法
CN112741598A (zh) * 2019-10-30 2021-05-04 株式会社理光 波形生成识别方法和计算机可读介质
CN111317500B (zh) * 2020-02-21 2021-05-07 孙磊 基于胎心胎动信号的智能穿戴系统
CN111265240A (zh) * 2020-02-21 2020-06-12 孙磊 胎心监护仪及胎心测量方法
CN111265239A (zh) * 2020-02-21 2020-06-12 孙磊 基于邻近计算的胎心检测信号处理及信息提取系统及方法
CN111227819B (zh) * 2020-02-21 2021-05-07 孙磊 多维通道传感器的胎心检测传感器矩阵的信号处理方法
CN111227820A (zh) * 2020-02-21 2020-06-05 孙磊 多维通道传感器的胎心检测传感器矩阵、胎心检测设备
CN116784861B (zh) * 2023-06-09 2024-03-29 中国科学技术大学 基于周期性快速独立向量分析的胎儿心电信号识别方法

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5372139A (en) 1991-06-24 1994-12-13 Paul Benjamin Crilly Method for suppressing a maternal electrocardiogram signal from a fetal electrocardiogram signal obtained with invasive and non-invasive techniques using an almost pure maternal electrocardiogram signal as a trigger
WO2003028550A2 (en) 2001-10-03 2003-04-10 Qinetiq Limited Apparatus for monitoring fetal heart-beat
WO2004084087A1 (en) 2003-03-22 2004-09-30 Qinetiq Limited Monitoring electrical muscular activity
WO2005039410A1 (en) 2003-10-14 2005-05-06 The University Of Nottingham Fetal surveillance
WO2005052848A2 (en) 2003-11-14 2005-06-09 Qinetiq Limited Dynamic blind signal separation
US20050267376A1 (en) 2004-05-28 2005-12-01 Dorothee Marossero Maternal-fetal monitoring system
US20070066908A1 (en) 2003-10-31 2007-03-22 Graupe Menachem H Separation of one or more fetal heart component signals from heart signal information obtained from a pregnant female
US20070233203A1 (en) * 2006-02-01 2007-10-04 Euliano Tammy Y Obstetric Analgesia System
EP1844706A1 (en) 2005-01-31 2007-10-17 Tohoku University Electrocardiogram signal processing method and electrocardiogram signal processing device
WO2009013246A1 (en) 2007-07-20 2009-01-29 Stichting Voor De Technische Wetenschappen Fetal monitoring

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA1328123C (en) * 1986-10-08 1994-03-29 Nigel John Randall Intrauterine probe

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5372139A (en) 1991-06-24 1994-12-13 Paul Benjamin Crilly Method for suppressing a maternal electrocardiogram signal from a fetal electrocardiogram signal obtained with invasive and non-invasive techniques using an almost pure maternal electrocardiogram signal as a trigger
WO2003028550A2 (en) 2001-10-03 2003-04-10 Qinetiq Limited Apparatus for monitoring fetal heart-beat
US20040243015A1 (en) * 2001-10-03 2004-12-02 Smith Mark John Apparatus for monitoring fetal heart-beat
WO2004084087A1 (en) 2003-03-22 2004-09-30 Qinetiq Limited Monitoring electrical muscular activity
WO2005039410A1 (en) 2003-10-14 2005-05-06 The University Of Nottingham Fetal surveillance
US20070066908A1 (en) 2003-10-31 2007-03-22 Graupe Menachem H Separation of one or more fetal heart component signals from heart signal information obtained from a pregnant female
WO2005052848A2 (en) 2003-11-14 2005-06-09 Qinetiq Limited Dynamic blind signal separation
US20050267376A1 (en) 2004-05-28 2005-12-01 Dorothee Marossero Maternal-fetal monitoring system
US20050267377A1 (en) 2004-05-28 2005-12-01 Dorothee Marossero Maternal-fetal monitoring system
WO2005117692A2 (en) 2004-05-28 2005-12-15 University Of Florida Research Foundation, Inc. Maternal-fetal monitoring system
EP1844706A1 (en) 2005-01-31 2007-10-17 Tohoku University Electrocardiogram signal processing method and electrocardiogram signal processing device
US20070233203A1 (en) * 2006-02-01 2007-10-04 Euliano Tammy Y Obstetric Analgesia System
WO2009013246A1 (en) 2007-07-20 2009-01-29 Stichting Voor De Technische Wetenschappen Fetal monitoring

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
GB Search Report dated Mar. 18, 2009.

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8897862B2 (en) 2011-01-25 2014-11-25 Public University Corporation Nara Medical University Fetal cardiac potential signal extraction program, fetal cardiac potential signal discriminating apparatus, and pregnancy monitoring system using the same
US8892181B2 (en) 2011-10-21 2014-11-18 Mindchild Medical, Inc. Non-invasive fetal monitoring
US20150335886A1 (en) * 2012-11-26 2015-11-26 Guangzhou Sunray Medical Apparatus Co., Ltd. Bio-feedback type device for physical pain relief at childbirth
US9333349B2 (en) * 2012-11-26 2016-05-10 Guangzhou Sunray Medical Apparatus Co., Ltd. Bio-feedback type device for physical pain relief at childbirth
WO2015038572A1 (en) * 2013-09-10 2015-03-19 Guerrero Juan R Diagnostic digital data mining of biological waves
US11224402B2 (en) 2017-09-13 2022-01-18 Koninklijke Philips N.V. Fetal heart rate detection method to discriminate from other periodic signal

Also Published As

Publication number Publication date
JP2009160410A (ja) 2009-07-23
DE102009003317A1 (de) 2009-07-09
US20090177101A1 (en) 2009-07-09
GB2456375B (en) 2012-05-23
GB2456375A (en) 2009-07-15
JP5271718B2 (ja) 2013-08-21
DE102009003317B4 (de) 2014-09-11
GB0823016D0 (en) 2009-01-28

Similar Documents

Publication Publication Date Title
US8064991B2 (en) Method of fetal and maternal ECG identification across multiple EPOCHS
US7809433B2 (en) Method and system for limiting interference in electroencephalographic signals
US7949389B2 (en) Fetal ECG monitoring
US9050007B1 (en) Extraction of cardiac signal data
Hassan et al. Combination of canonical correlation analysis and empirical mode decomposition applied to denoising the labor electrohysterogram
US8892181B2 (en) Non-invasive fetal monitoring
US7509159B2 (en) Method and apparatus for detecting cardiac repolarization abnormality
EP2428160A2 (en) Monitoring physiological activity using partial state space reconstruction
Mak et al. An automated ECG-artifact removal method for trunk muscle surface EMG recordings
US20130102856A1 (en) Non-invasive detection of fetal or maternal illness
EP2016894A1 (en) ECG signal processing
US7831300B2 (en) Extrapolating ICA knowledge from one epoch to another for improved fetal ECG separation
John et al. Extraction of foetal ECG from abdominal ECG by nonlinear transformation and estimations
US7197358B2 (en) Identifying infants at risk for sudden infant death syndrome
Vrins et al. Abdominal electrodes analysis by statistical processing for fetal electrocardiogram extraction
US20190374123A1 (en) Predicting atrial fibrillation or stroke using p-wave analysis
Rampil Elements of EEG signal processing
CN112022151A (zh) 一种脑电棘慢波的处理及识别方法
Balaji et al. Revisiting Derivative Based Methods on QRS Detections from an ECG Signal
Widatalla et al. Detection of End of T-Wave in Fetal ECG Using Recurrence Plots
Talib et al. Enhancement Separation of ECG Signals for Twin Fetuses Based on Modified Blind Source Separation
Pani et al. A real-time algorithm for tracking of foetal ECG sources obtained by block-on-line BSS techniques
RU2791006C1 (ru) Система и способ автоматизированного анализа и интерпретации электрокардиограммы
Anisha et al. Recognition and eradication of prime artefact from the abdominal electrocardiogram
Kamiński et al. ECG signal processing for deceleration capacity assessment

Legal Events

Date Code Title Description
AS Assignment

Owner name: THE GENERAL ELECTRIC COMPANY, NEW YORK

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HERSH, LAWRENCE T.;KOLLURI, SAI;FRIEDMAN, BRUCE A.;REEL/FRAME:021508/0413

Effective date: 20071218

REMI Maintenance fee reminder mailed
LAPS Lapse for failure to pay maintenance fees
STCH Information on status: patent discontinuation

Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362

FP Lapsed due to failure to pay maintenance fee

Effective date: 20151122